As we have became the fattest nation in the world, obesity is an increasingly important cause of myocardial morbidity and mortality in the United States. Obesity indirectly contributes to heart disease by increasing plasma lipids and predisposing to diabetes and hypertension (i.e. traditional cardiovascular risk factors). In addition, recent work from Roger Unger advances the novel hypothesis that obesity per se constitutes a direct cause of left systolic ventricular dysfunction and hypertrophy by promoting cardiac steatosis. Major Hypothesis: Cardiac steatosis is an integral feature of human obesity, contributing to decrease of LV systolic function. I further hypothesize that these functional abnormalities of the human heart are provoked by the development of non-insulin dependent diabetes mellitus (type 2 diabetes) and can be reversed by treatment with Thiazolidinediones. Specific Aims: In human subjects without heart disease, I will measure lipid deposition in the ventricular septum using double gated localized proton nuclear magnetic resonance (1H NMR) as well as LV systolic function with magnetic resonance imaging (MRI) to accomplish the following specific aims: Aim 1: To document the intra-subject reproducibility of the 1H NMR measurement of intra-myocardial lipid content. Aim 2: To establish the relation between adiposity and myocardial lipid deposition over a wide range of BMI and determine the impact of gender, and ethnicity on these relations. Aim 3: To determine if intra-myocardial lipid deposition is greater in individuals with IGT and insulin resistance than in these with normal glucose tolerance matched for BMI, age, gender and ethnicity. I hypothesize that the greater myocardial lipid in diabetic and prediabetic individuals will be accompanied by decreased LV systolic function. Aim 4: To perform randomized prospective study to test the hypothesis that the elevated myocardial lipid content and decreased systolic LV function can be reversed or minimized by the treatment with Thiazolidinediones. In contrast, I predict that functional cardiac abnormalities will be unaffected when diabetic subjects are treated with sulfonylureases that have no effect on PPAR-gamma.